Inter-Domain Point-To-Multipoint Path Computation in a Computer Network

1. A method, comprising: receiving, at a path computation element (PCE) of a local domain, a path computation request for an inter-domain point-to-multipoint (P2MP) path to a plurality of destinations; receiving, from one or more adjacent predecessor domains, a cost associated with using one or more local ingress border routers of the local domain for the P2MP path; for each destination in the local domain, i) computing a local portion of the P2MP path from one or more of the local ingress border routers to each destination in the local domain, and ii) returning a list of local ingress border routers used in the P2MP path for each destination in the local domain to each predecessor domain; and for each destination not in the local domain, i) computing a local portion of the P2MP path from one or more of the local ingress border routers to one or more successor ingress border routers of one or more adjacent successor domains, ii) computing a corresponding cost to use the successor ingress border routers, iii) informing a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domains, iv) receiving a list of successor ingress border routers used in the P2MP path from each successor domain, v) pruning segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list for the respective successor domain, and vi) for each predecessor domain, returning a list of local ingress border routers used in the P2MP path once pruning is performed to the respective predecessor domain.

2. The method as in claim 1 , further comprising: sending the computed local portion of the P2MP path without pruned segments to a root PCE.

3. The method as in claim 1 , wherein the computed local portion of the P2MP path is a pass-through path.

4. The method as in claim 1 , further comprising: constructing an augmented topology based on predecessor domains upon which to compute the local portion of the P2MP path from one or more of the local ingress border routers to one or more successor ingress border routers of the one or more adjacent successor domains.

5. The method as in claim 1 , further comprising: applying a weight to the computed corresponding cost.

6. The method as in claim 1 , wherein the path computation request is for diverse inter-domain P2MP paths.

7. The method as in claim 1 , further comprising: ensuring that the computed local portion of the P2MP path without pruned segments is used for a signaled tunnel over the P2MP path.

8. The method as in claim 7 , wherein the step of ensuring further comprises: utilizing encrypted route objects (EROs).

9. The method as in claim 7 , wherein the step of ensuring further comprises: utilizing loose hops and path keys.

10. The method as in claim 1 , further comprising: receiving within the path computation request a set of path destinations within the local domain, a set of predecessor domains, and a set of successor domains.

11. A method, comprising: receiving, at a path computation element (PCE) of a root domain, a path computation request for an inter-domain point-to-multipoint (P2MP) path from a source in the root domain to a plurality of destinations; for each destination in a local domain, computing a local portion of the P2MP path to each destination in the local domain; and for each destination not in the local domain, i) computing a local portion of the P2MP path to one or more successor ingress border routers of one or more adjacent successor domains, ii) computing a corresponding cost to use the successor ingress border routers, iii) informing a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domains, iv) receiving a list of successor ingress border routers used in the P2MP path from each successor domain, and v) pruning segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list for the respective successor domain.

12. The method as in claim 11 , further comprising: computing a domain-level directed acyclic graph (DAG) from the source to the destinations.

13. The method as in claim 12 , further comprising: sending a P2MP path computation request to PCEs of the domains in the DAG.

14. The method as in claim 12 , wherein the DAG excludes domains based on policy.

15. A method, comprising: receiving, at a path computation element (PCE) of a root domain, a path computation request for an inter-domain point-to-multipoint (P2MP) path from a source in the root domain to a plurality of destinations; for each destination in the root domain, computing a local portion of the P2MP path from the source to each destination in the root domain; and for each destination not in the root domain, i) computing a local portion of the P2MP path from the source to one or more successor ingress border routers of one or more adjacent successor domains, ii) computing a corresponding cost to use the successor ingress border routers, iii) informing a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domains, iv) receiving a list of successor ingress border routers used in the P2MP path from each successor domain, and v) pruning segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list for the respective successor domain.

16. The method as in claim 15 , further comprising: receiving one or more computed local portions of the P2MP path without pruned segments.

17. The method as in claim 16 , further comprising: combining the one or more computed local portions of the P2MP path without pruned segments into the inter-domain P2MP path.

18. The method as in claim 17 , further comprising: returning the inter-domain P2MP path to a path computation client (PCC).

19. The method as in claim 17 , further comprising: establishing a P2MP tunnel over the inter-domain P2MP path.

20. A path computation element (PCE), comprising: one or more network interfaces; one or more processors coupled to the network interfaces and adapted to execute one or more processes; and a memory adapted to store an inter-domain point-to-multipoint (P2MP) path computation process executable by each processor, the path computation process when executed operable to: i) receive a path computation request for an inter-domain P2MP path to a plurality of destinations; ii) receive, from one or more adjacent predecessor domains, a cost associated with using one or more local ingress border routers of the local domain for the P2MP path; iii) for each destination in the local domain, a) compute a local portion of the P2MP path from one or more of the local ingress border routers to each destination in the local domain, and b) return a list of local ingress border routers used in the P2MP path for each destination in the local domain to each predecessor domain; and iv) for each destination not in the local domain, a) compute a local portion of the P2MP path from one or more of the local ingress border routers to one or more successor ingress border routers of one or more adjacent successor domains, b) compute a corresponding cost to use the successor ingress border routers, c) inform a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domains, d) receive a list of successor ingress border routers used in the P2MP path from each successor domain, e) prune segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list for the respective successor domain, and f) for each predecessor domain, return a list of local ingress border routers used in the P2MP path once pruning is performed to the respective predecessor domain.

21. The PCE as in claim 20 , wherein the path computation process when executed is further operable to send the computed local portion of the P2MP path without pruned segments to a root PCE.

22. The PCE as in claim 20 , wherein the computed local portion of the P2MP path is a pass-through path.

23. A path computation element (PCE), comprising: one or more network interfaces; one or more processors coupled to the network interfaces and configured to execute one or more processes; and a memory configured to store a path computation process executable by the one or more processors, the path computation process when executed to receive a path computation request for an inter-domain point-to-multipoint (P2MP) path to extend from a source in a root domain to a plurality of destinations, for each destination in the root domain, compute a local portion of the P2MP path from the source to each destination in the root domain, and for each destination not in the root domain, compute a local portion of the P2MP path from the source to one or more successor ingress border routers of one or more adjacent successor domains, compute a corresponding cost to use the successor ingress border routers, inform a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domain, receive a list of successor ingress border routers used in the P2MP path, and prune segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list.

24. The PCE as in claim 23 , wherein the path computation process when executed is further to compute a domain-level directed acyclic graph (DAG) from the source to the destinations.

25. The PCE as in claim 24 , wherein the path computation process when executed is further to send a start message to one or more PCEs of the one or more domains in the DAG.

26. The PCE as in claim 23 , wherein the path computation process when executed is further to receive one or more computed local portions of the P2MP path without pruned segments.

27. The PCE as in claim 26 , wherein the path computation process when executed is further to combine the one or more computed local portions of the P2MP path without pruned segments into the inter-domain P2MP path.

28. The PCE as in claim 27 , wherein the path computation process when executed is further to return the inter-domain P2MP path to a path computation client (PCC).

29. A path computation element (PCE), comprising: means for interfacing with a network; means for executing instructions; and means for storing instructions that when executed: receives a path computation request for an inter-domain point-to-multipoint (P2MP) path to extend from a source in a root domain to a plurality of destinations, for each destination in the root domain, compute a local portion of the P2MP path from the source to each destination in the root domain, and for each destination not in the root domain, compute a local portion of the P2MP path from the source to one or more successor ingress border routers of one or more adjacent successor domains, compute a corresponding cost to use the successor ingress border routers, inform a PCE of each successor domain of the cost associated with using the successor ingress border routers of the respective successive domain, receive a list of successor ingress border routers used in the P2MP path, and prune segments of the computed local portion of the P2MP path that lead to successor ingress border routers not in the list.